Good Practice guide to measure roundness on …

1 good Practice guide to measure roundness on roller machines and to estimate their uncertainty Bj rn Hemming, VTT Technical Research Centre of Finland Ltd, Finland Thomas Widmaier, Aalto University School of Engineering, Finland We gratefully acknowledge the funding from the European Metrology Research Programme (EMRP). The EMRP is jointly funded by the EMRP participating countries within EURAMET and the European Union. IND62 TIM Traceable in-process dimensional measurement Page 2 of 20 Table of Contents List of Figures .. 3 1 Introduction .. 4 2 Determination of roundness measurement uncertainties using the R-MS .. 4 3 Determination of roundness measurement uncertainties using Monte Carlo simulation .. 7 Four-point roll roundness measurement .. 7 Two-point method .. 7 Three-point Ozono roundness measurement method.

2 8 Hybrid four-point roundness measurement .. 9 Probes .. 10 Uncertainty evaluation by simulation .. 11 roundness profile used in simulation .. 11 Probability distributions .. 12 4 Conclusion .. 18 5 References .. 19 6 Appendix 1. Acceptance test.. 20 IND62 TIM Traceable in-process dimensional measurement Page 3 of 20 List of Figures Figure 1. Typical orientation of probes in a four point measurement systems .. 4 Figure 2. Setup for the measurement of roundness , cylindricity and diameter using a multi-point measurement device.. 5 Figure 3. Polar plot of disc type C.. 6 Figure 4. Comparison of the measurement results of the harmonics of the calibration disc type C made on three different sites.. 6 Figure 5. Two-point measurement transducer orientations and locations of a round 7 Figure 6. Modified saddle-type roll calipers for two-point measurement.

3 A) With fixed arms. B) With swing arms.. 7 Figure 7. Two point measurement cannot be used to measure odd lobe geometry errors, because the method suffers from harmonic filtration, , the method cannot separate the odd lobe geometry error from the error motion of the rotating axis.. 8 Figure 8. Orientation and location of the three run-out measurement transducers of the Ozono method.. 8 Figure 9. Orientation of probes [S1-S4] in a four-point measurement system when measuring a round 10 Figure 10. Principle of the hybrid four-point method.. 10 Figure 11. Measurement rod with MT12 probe.. 11 Figure 12. Calibration disc.. 11 Figure 13. a) Designed test profile with 2 30 UPR. b) Designed wave amplitudes of the test profile. 12 Figure 14. Alignment/position error of a rod.. 14 Figure 15. Output from Monte Carlo simulation (N = 10 000) where the different standard uncertainties are shown as error bars.

4 This simulation was run without any error sources to get a result about the ideal amplitude values.. 15 Figure 16. Output from Monte Carlo simulation (N = 10 000) where the different standard uncertainties are shown as error bars. This simulation was run with all the specified error sources from Table 1.. 15 Figure 17. Output from Monte Carlo simulation (N = 10 000) where the different standard uncertainties are shown as error bars. This simulation was run with thermal error sources from Table 2 ( =1 C).. 16 Figure 18. Outputs from four Monte Carlo simulations (N = 10 000) where the different standard uncertainties are shown as error bars. These simulations were run with the probe error value from Table 2 ( =1 m).. 17 Figure 19. Outputs from four Monte Carlo simulations (N = 10 000) where the different standard uncertainties are shown as error bars.

5 These simulations were run with the rod alignment error ( =1 ).. 17 IND62 TIM Traceable in-process dimensional measurement Page 4 of 20 1 Introduction Large scale rotors in the paper and steel industry are called rolls. These rolls are periodically reground and roundness measurements are made throughout the machining process. Dimensional measurement systems for large rolls (diameter < 2000 mm) are available on the market. 2 Determination of roundness measurement uncertainties using the R-MS The measurement standards are intended to quantify error sources found in three point measurement systems and in four point measurement systems (Figure 1). The developed measurement standards are also useful for calibration of two-point and one-point measurement systems. The name in process roundness measurement systems will be used for these all categories in this guide .

6 Figure 1. Typical orientation of probes in a four point measurement systems All standards are discs with the diameter of 500 mm 550 mm. This is the largest diameter that can be easily measured both in laboratories and not too small to be measured by in process roundness measurement systems in industry. The thickness of the discs will be 30 mm 50 mm. The requirements are shown in table 1. Table 1. Requirements for the measurement standards (R-MS). Name Form roundness error / m R-MS, type A round 0 m 2 m R-MS, type B 21 UPR 20 m 25 m R-MS, type C extended multiwave, 2 UPR 30 UPR 10 m / undulation IND62 TIM Traceable in-process dimensional measurement Page 5 of 20 The type A standard is almost perfectly round, still not too expensive to manufacture. With a roundness error below 2 m this standard helps to reveal errors like noise and thermal drift.

7 Type B is selected as it has one characteristic form of a 21 UPR wave. The propagation of error of single probes at in process roundness measurement systems is expected to be revealed by this disc. The type C, extended multi wave, consists of several waves. Standards of this type have previously been used and they are expected to work as overall test standard. All standards are calibrated by roundness measuring instrument with rotary table. The measurement standards are intended to quantify error sources found in three point measurement systems and in four point measurement systems (see Figure 2). Figure 2. Setup for the measurement of roundness , cylindricity and diameter using a multi-point measurement device. IND62 TIM Traceable in-process dimensional measurement Page 6 of 20 In Figure 3 is shown a polar plot of disc C measured using a roundness instrument at MIKES.

8 Additional tests with the manufactured calibration disc type C were carried out on three sites: with the four-point roll geometry measurement device (RollCal 3) at an industrial partner (IP), with the four-point measurement device (own model) at Aalto University and with the CMM (Legex 9106) at MIKES (Figure 4). Figure 3. Polar plot of disc type C. Figure 4. Comparison of the measurement results of the harmonics of the calibration disc type C made on three different sites. In appendix 1 an Acceptance test for roll grinding machine is given. IND62 TIM Traceable in-process dimensional measurement Page 7 of 20 3 Determination of roundness measurement uncertainties using Monte Carlo simulation Four-point roll roundness measurement The four-point roundness measurement method is a combination of the two-point method and the Ozono three-point method.

9 Thus they both are briefly discussed here. Two-point method The two-point method uses only two sensors. For some applications the other sensor can be replaced by a fixed point, see Figure 5 right hand side. Practical implementations of this kind of measurement devices are modified roll mikes or roll calipers, see Figure 6. Figure 5. Two-point measurement transducer orientations and locations of a round workpiece. Figure 6. Modified saddle-type roll calipers for two-point measurement. A) With fixed arms. B) With swing arms. This method measures the diameter profile or diameter variation profile. In principle, the only difference between the diameter variation profile and diameter profile is that the from the variation profile the average or the minimum diameter value has been subtracted. The diameter variation measurement is commonly used in large roll grinding machine .

10 There the measured profile is often called roundness profile , although two-point measuring method IND62 TIM Traceable in-process dimensional measurement Page 8 of 20 cannot measure the true roundness profile, because it suffers from a harmonic filtration, as shown in Figure 7, where a Relaux triangle is illustrated. Using this type of diameter measuring device one cannot measure odd lobe shapes like triangular, 5-lobe, 7-lobe etc. geometries, because the method is unable to separate the geometry error of the cross-section from the error motion of the rotating axis. Figure 7. Two point measurement cannot be used to measure odd lobe geometry errors, because the method suffers from harmonic filtration, , the method cannot separate the odd lobe geometry error from the error motion of the rotating axis.